order flow and liquidity around nyse trading halts · 3 news-related trading halts are also...

Order Flow and Liquidity around NYSETrading Halts

SHANE A. CORWIN and MARC L. LIPSON*

ABSTRACT

We study order f low and liquidity around NYSE trading halts. We find that mar-ket and limit order submissions and cancellations increase significantly duringtrading halts, that a large proportion of the limit order book at the reopen is com-posed of orders submitted during the halt, and that the market-clearing price atthe reopen is a good predictor of future prices. Depth near the quotes is unusuallylow around trading halts, though specialists and0or f loor traders appear to provideadditional liquidity at these times. Finally, specialists appear to “spread the quote”prior to imbalance halts to convey information to market participants.

THE COSTS AND BENEFITS ASSOCIATED with New York Stock Exchange tradinghalts are the subject of continuous debate. The stated purpose of tradinghalts is to allow investors a chance to react to new information and to facilitatethe orderly emergence of a new equilibrium price ~see, e.g., NYSE ~1999!!.However, little is known about the extent or nature of investor reactions totrading halts. Furthermore, it is unclear whether these reactions support orconfound attempts to reach a new equilibrium price. To address these issues,we use nonpublic NYSE data on all SuperDOT ~electronic! orders to analyzethe effects of trading halts on order f low, spreads, and the limit order book.We also examine the relationships between the limit order book and thereopening price and between liquidity and post-halt volatility.

Information on order f low is an important missing element in the ongoingdebate about the usefulness of trading halts. One argument against tradinghalts is that they impede price formation because trading aggregates infor-mation that is distributed across market participants. However, submis-sions and cancellations of orders during a halt may provide an alternativesource of information that, combined with published indications of possiblereopening prices, may assist with price formation. An argument in favor oftrading halts is that when traders are given an opportunity to cancel orders

* Corwin and Lipson are from the Terry College of Business at the University of Georgia. Wethank Franklin Allen, Jeff Bacidore, Michael Goldstein, Jeff Harris, Ken Kavajecz, Jeff Netter,George Sofianos, Dan Weaver, and seminar participants at the Federal Reserve Bank of Atlanta,the University of Minnesota, Vanderbilt University, the University of Virginia, and the NYSEJB Seminar for helpful comments. We are also grateful to the New York Stock Exchange forproviding data. The comments and opinions expressed in this paper are the authors’ and do notnecessarily ref lect those of the directors, members, or officers of the New York Stock Exchange,Inc. The authors are responsible for any errors.

THE JOURNAL OF FINANCE • VOL. LV, NO. 4 • AUGUST 2000

1771

during extreme market changes, they may be more willing to supply liquid-ity during normal conditions. Our analysis allows us to examine whethertraders take advantage of this opportunity to reposition their trading interest.

Whereas analysis of order f low provides insight into the actions of tradersduring halts, the net effect of this activity on liquidity is ref lected in spreadsand the limit order book. The limit order book competes with specialists andthe trading f loor for order f low and provides a pool of trading interest thatcan dampen temporary price f luctuations. Thus, changes in the limit orderbook may impact spreads and volatility. We characterize the limit order bookand bid-ask spreads before, during, and after trading halts to provide newevidence about changes in liquidity around halts and the impact of the limitorder book on volatility and price discovery around halts.1

We find that limit order submissions and cancellations are extremely highduring halts and remain high for several hours after the halt. In addition,we find that there is a surge in market order submissions and cancellationsduring halts and just before order imbalance halts. Although the cancella-tion of orders might be expected as traders wait to see how market condi-tions unfold, the large increase in submissions suggests that they are alsowilling to reenter the market before trading resumes. Furthermore, changesin order f low prior to order imbalance halts are limited to the 30 minutesimmediately preceding the halt. This result suggests that the change in mar-ket conditions is abrupt and there is little time for the NYSE specialist todraw liquidity to the market.

We find that limit order book depth near the quotes is unusually low before,during, and after trading halts. This reduction likely ref lects investors’ re-duced willingness to supply liquidity at these times. The significant lack of depthjust prior to a trading halt is also consistent with halts being called when thereis valid concern about the quality of current prices. Notably, we find that a largeproportion of the limit order book at the reopen is composed of orders that wereplaced during the trading halt. Combined with the increased submissions andcancellations observed during halts, this result suggests that traders use thehalt as a chance to reposition their trading interest. As further evidence of thisrepositioning, we calculate the price that clears supply and demand in the limitorder book just prior to the reopen. Our tests show that this market-clearingprice is a good predictor of subsequent equilibrium prices.

In examining bid-ask spreads, we find that the quoted spread and severalmeasures of limit order book spread are unusually high at the reopen. Thisincrease generally dissipates within two minutes for order imbalance haltsand within 30 minutes for news halts. Interestingly, we find evidence of adramatic increase in quoted spreads prior to order imbalance halts. Meanquoted spreads increase from approximately $0.23 to over $1.80 in the 10 to15 minutes prior to imbalance halts. This increase is consistent with the specialist

1This analysis focuses on the net effect of order f low on the limit order book during times ofmarket stress. See Biais, Hillion, and Spatt ~1995! for an analysis of the dynamic interactionsbetween order f low and the limit order book during normal market conditions.

1772 The Journal of Finance

“spreading the quote” in response to a substantial order imbalance. By spread-ing the quote, the specialist conveys information about the imbalance to mar-ket participants and gives market orders a chance to be canceled. Consistentwith the decrease in limit-book depth, we also find that the specialist’s quoteddepth is unusually low prior to and immediately after trading halts. However,the proportion of depth contributed by f loor participants actually increases im-mediately after halts, suggesting that f loor participants, including the spe-cialist, step in to provide liquidity after trading halts.

Consistent with Lee, Ready, and Seguin ~1994!, we find that both volumeand volatility increase significantly after NYSE trading halts. We also findweak evidence that this increased volatility is related to decreased depth inthe limit order book. However, changes in liquidity appear to explain, atbest, a small portion of the increased volatility observed after trading halts.

It should be emphasized that our results cannot resolve the question ofwhether trading halts and related actions like “spreading the quote” aredesirable. The central problem is that we cannot know what would haveoccurred in the absence of a halt or what equilibrium trading patterns wouldbe in a market where halts are not permitted. However, our results do pro-vide a more complete understanding of trading activity during halts and theassociated effects on liquidity, price discovery, and volatility.

The remainder of the paper is organized as follows. In Section I, we dis-cuss the central arguments associated with trading halts and the relatedevidence. The trading halt sample is described in Section II. In Section III,we analyze order f low and the state of the limit order book around tradinghalts. In Section IV, we examine the role of the limit order book in deter-mining posthalt prices. Section V presents our analysis of posthalt volatility,and Section VI concludes.

I. Trading Halts

There are two primary categories of security-specific trading halts on theNYSE: news halts and order imbalance halts.2 News halts include both newspending and news dissemination halts and are initiated by exchange offi-cials when an information release is expected to have a significant impacton prices.3 Order imbalance halts are initiated by the exchange specialistwhen a large imbalance exists between buy and sell orders. During bothtypes of halts, the specialist posts quote indications to signal the possiblereopening price, and traders are permitted to submit or cancel orders. In

2 Our analysis and discussion focus exclusively on individual-security halts. NYSE Rule 80Balso allows for marketwide trading halts ~or circuit breakers! in the event of extremely largemarket movements. These circuit breakers were initiated for the first and only time on October27, 1997. Goldstein and Kavajecz ~2000b! provide a detailed analysis of liquidity provision dur-ing this marketwide circuit breaker.

3 News-related trading halts are also implemented on the Nasdaq stock market, though nomechanism exists on that exchange for order imbalance halts. See Christie et al. ~2000! for adiscussion and analysis of Nasdaq trading halts.

Order Flow and Liquidity around NYSE Trading Halts 1773

addition, whereas the specialist is expected to provide price discovery duringthe halt, the specialist’s obligations to provide price continuity and pricestabilization are suspended until trading resumes.4 Trading is reopened witha call market similar to that used at the morning open.

Several papers delineate the benefits of trading halts. Greenwald and Stein~1988, 1991! argue that traders are reluctant to enter a market when theyare unsure of the price at which their market orders will be executed. Intheir model, trading halts can facilitate price discovery by reducing this“transactional risk” and thereby encouraging traders to participate in themarket. Kodres and O’Brien ~1994! highlight an additional benefit of trad-ing halts based on the fact that traders may not be able to instantaneouslyadjust their trading interests to new information ~due to technological lim-itations or the costs of continuous monitoring!. They show that Pareto-optimal risk sharing can be achieved through the use of price limits thatprevent sudden, extreme movements in prices.

In contrast, a number of models suggest that trading halts may actuallyreduce the informativeness of prices. In these models, trading is necessaryfor information distributed across multiple participants to be ref lected inprices ~see, e.g., Brown and Jennings ~1989! and Grundy and McNichols~1989!!. Consistent with these models, Amihud and Mendelson ~1987!, Stolland Whaley ~1990!, and Gerety and Mulherin ~1994! find that market open-ings are associated with high levels of temporary volatility. In addition, Leeet al. ~1994! find evidence of unusually high volatility after NYSE tradinghalts, and Christie, Corwin, and Harris ~2000! document unusually highvolatility after Nasdaq trading halts. Grossman ~1990! expresses the moredirect concern that a trading halt “merely prevents consenting adults fromcarrying out their desires on the f loor of the stock exchange.”

This paper contributes to the trading halt literature by providing some ofthe first empirical evidence related to the actions of investors during tradinghalts. In particular, we document that traders do, in fact, adjust their posi-tions during trading halts. This activity is a necessary condition for a re-duction in transactional risk. In addition, although overall liquidity is reducedduring this period of market stress, we show that submissions and cancel-lations of orders during halts result in significant price discovery.

II. Data and Sample Characteristics

Using the NYSE’s trading halt files, we identified 469 intraday trading haltscalled for NYSE-listed common stocks during 1995 and 1996. Because we wishto analyze trading activity and the state of liquidity before and after halts,

4 The rules governing specialist trading activities generally prevent specialists from makingtrades that reinforce market movements ~destabilizing trades!. In addition, specialists are ex-pected to maintain price continuity by limiting price changes larger than the minimum ticksize. The combination of these two obligations essentially forces the specialist to buy as pricesfall and sell as prices rise.


delayed openings are not considered. For similar reasons, we exclude halts calledbefore 10:00 a.m. ~44 halts! and cases where the same stock experienced morethan one halt on the same day ~14 halts!. Finally, we exclude halts that werenot resolved within one trading day ~104 halts! and halts for which we haveincomplete data ~12 halts!. This leaves a sample of 295 trading halts.

Data on each trading halt are obtained from the NYSE’s trading halt files.These data include the time of the halt, the halt type, the last trade priceprior to the halt, the reopening price, the reopening time, and the volumetraded in the reopening transaction. In addition, we collect trade and quotedata for each halt, in addition to NYSE System Order ~SOD! data for the21-day period from 10 days prior to the halt through 10 days after the halt.The trade and quote data come from the NYSE’s Trade and Quote ~TAQ!database. The SOD data include detailed information on all SuperDOT ~elec-tronic! orders. This information includes the order submission time, the num-ber of shares submitted, whether the order is a buy or a sell, the order type~market or limit!, order conditions ~day order, good-till-canceled, etc.!, andthe limit price ~if applicable!. If any portion of the order is executed or can-celed, the data also include the time of the execution or cancellation, thenumber of shares executed or canceled, and the execution price ~if applicable!.

Summary statistics for the trading halt sample are provided in Table I. Ofthe 295 halts in the sample, 44 are order imbalance halts and 251 are newshalts. The mean value of trading halt duration is 81.4 minutes in the full

Table I

Summary Statistics for Trading HaltsThe sample is taken from the set of all news and order imbalance halts called for NYSE-listedcommon stocks during 1995 and 1996. We exclude opening delays, halts prior to 10:00 a.m.,halts that are not resolved prior to the close, halts for which we do not have complete data, andobservations for which two or more halts are called on the same day for a particular stock. Pricechange is measured from the last trade price prior to the halt to the reopening trade price.Duration is defined as the number of minutes from the beginning of the halt to the reopeningtrade.

Mean Values

Sample

Numberof

HaltsDuration

in MinutesReopening

Volume

PriceChange

~%!

AbsolutePrice Change

~%!

Full sample 295 81.41 66,294 0.47 3.93

Order imbalance halts 44 55.71 167,147 1.16 5.71Positive price change 21 63.53 195,962 7.19 7.19Zero price change 5 49.85 55,240 0.00 0.00Negative price change 18 48.22 164,617 25.57 5.57

News halts 251 85.92 48,614 0.35 3.62Positive price change 90 89.93 61,704 5.54 5.54Zero price change 52 83.03 21,654 0.00 0.00Negative price change 109 83.98 50,668 23.76 3.76


sample and is higher for news halts ~85.9 minutes! than for order imbalancehalts ~55.7 minutes!. The average absolute return during trading halts is3.93 percent and is smaller for news halts ~3.62 percent! than for order im-balance halts ~5.71 percent!. Twenty-one of the 44 order imbalance haltsresulted in positive price changes, five led to no price change, and 18 led tonegative price changes. Ninety of the 251 news halts led to positive pricechanges, while 52 led to no price change and 109 resulted in negative pricechanges. These values are generally consistent with those reported in Leeet al. ~1994! for trading halts during 1988 and in Corwin ~1999! for tradinghalts during 1992.

III. Order Flow and the Limit Order Book

In this section, we examine the behavior of investors around NYSE trad-ing halts and the resulting effects on liquidity. We begin by analyzing sub-missions, executions, and cancellations of limit orders on days when haltsoccur. We then examine the effects of this activity on the condition of thelimit order book, bid-ask spreads, and quoted depth before and after thehalt.

A. Order Flow

Throughout the paper, we compare trading activity and market conditionson the halt day to the same characteristics on normal trading days. For allof the characteristics that we examine, we follow Lee et al. ~1994! in calcu-lating abnormal measures on the halt day relative to nonhalt days, stated asa percentage of the nonhalt day value. For example, for each time period wedefine abnormal order submissions on the halt day as:

100 * F Shares Submitted on the Halt Day 2 Mean Shares Submitted on Nonhalt Days

Mean Shares Submitted on Nonhalt Days G,

~1!

where we define nonhalt days as the 10 days prior to and the 10 days afterthe halt day; time periods on both halt and nonhalt days are defined basedon the halt time and reopening time. Throughout the paper, we report testsbased on the median abnormal statistic across all sample halts. Except wherenoted, results based on mean abnormal statistics are similar.

Submissions and cancellations of limit and market orders on the halt dayare summarized in Table II. For each period, the table lists the mediannumber of shares submitted, executed, or canceled and the median value ofthe abnormal statistic. The abnormal statistic can be interpreted as the per-centage difference between the halt day and the average nonhalt day. Forexample, the value of 236.28 reported in the fourth column of Table II sug-gests that market order submissions during period 24 ~defined below! are36.3 percent lower on the halt day than on the average nonhalt day.


Table II

Total Submissions, Executions, and Cancellations around Trading HaltsThe sample includes 295 order imbalance and news halts of NYSE-listed common stocks that were called after 10:00 a.m. Submissions, execu-tions, and cancellations are identified using NYSE System Order Data. The table lists the median number of shares submitted, executed, orcanceled and the median value of abnormal submissions, executions, and cancellations. The abnormal submissions measure is defined as theshares submitted on the halt day minus the mean number of shares submitted on nonhalt days, stated as a percentage of the nonhalt day mean.Abnormal cancellations are defined similarly. Nonhalt days include the 10 days before and 10 days after the halt. For the abnormal measures,***, **, and * indicate that the median is statistically significant at the 1, 5, and 10 percent levels, respectively, based on a signed-rank test. Thehalt period extends from the beginning of the halt to the reopening trade. Thirty-minute prehalt periods are measured backward from thebeginning of the halt. Thirty-minute posthalt periods are measured forward from the reopening time.

Market Order Submissions Market Order Cancellations Limit Order Submissions Limit Order Cancellations

Period N

HaltDay

Submissions

AbnormalSubmissions

~%!

HaltDay

Cancellations

AbnormalCancellations

~%!

HaltDay

Submissions

AbnormalSubmissions

~%!

HaltDay

Cancellations

AbnormalCancellations

~%!

Panel A: News Halts

24 182 880 236.28 0 −100.00*** 6,500 243.74 1,000 −66.47**23 197 500 −65.07*** 0 −100.00*** 5,000 233.43 812 −65.34*22 214 500 231.22 0 −100.00*** 4,536 239.19 1,000 −73.95*21 251 884 231.24 0 −100.00*** 6,200 226.65 2,000 245.02

During halt 251 4,227 56.82*** 500 618.88*** 41,340 150.45*** 38,900 371.50***11 222 6,585 689.84*** 0 2100.00 37,858 488.17*** 16,820 471.23***12 184 2,850 167.37*** 0 −100.00** 21,200 195.17*** 6,648 138.57***13 160 1,706 53.66*** 0 −100.00*** 18,482 129.04*** 5,030 23.32***14 119 1,630 22.98*** 0 −100.00*** 13,365 81.50*** 4,200 30.29***

Panel B: Order Imbalance Halts

24 24 3,232 21.42 0 2100.00 8,199 28.44 5,338 238.8723 31 900 230.00 0 −100.00*** 11,975 219.52 6,512 219.1822 37 850 242.15 0 −100.00*** 11,314 6.48 1,600 260.9821 44 13,009 569.51*** 50 189.86** 34,853 326.66*** 13,300 356.26***

During halt 44 11,677 319.71*** 9,705 8,028.89*** 116,452 1,183.85*** 78,240 1,847.19***11 40 18,956 1,484.22*** 0 2100.00 54,156 815.63*** 34,995 1,298.63***12 37 9,767 204.51*** 0 2100.00 28,080 231.53*** 12,350 335.54***13 36 6,675 340.45*** 0 2100.00 25,831 160.79*** 13,341 234.22***14 31 4,144 87.90*** 0 2100.00 17,000 126.98*** 7,900 69.64***

Ord

erF

lowan

dL

iquid

ityarou

nd

NY

SE

Trad

ing

Halts

1777

The halt period begins at the time the halt is called and includes all ac-tivity through the reopening trade. Thirty-minute prehalt periods are mea-sured backward from the halt beginning time to the open, and 30-minuteposthalt periods are measured forward from the reopening time to the close.We include results for the period from two hours prior to the halt to twohours after the halt. In addition, we report separate results for news halts~Panel A! and order imbalance halts ~Panel B!. Statistical tests are based onthe signed-rank test.

Panel A of Table II lists results for the sample of news-related halts. In thefour pre-halt periods ~24 through 21!, measures of abnormal trading activ-ity ~i.e., submissions and cancellations! are generally negative. However, thereis little evidence of statistically reliable differences.5 These results suggestthat there is little change in prehalt activity relative to nonhalt days. Be-ginning at the halt time, however, trading activity changes dramatically rel-ative to nonhalt days. During the halt, submissions and cancellations aresignificantly higher than nonhalt day values for both market and limit orders.Most notably, limit order cancellations during the halt period are 371 per-cent higher than nonhalt day values, and market order cancellations areover 600 percent higher. In addition, market order submissions during thehalt increase by 50 percent relative to nonhalt days, and limit order sub-missions more than double ~an increase of 150 percent!. Whereas marketorder cancellations are only evident during the halt, other categories of trad-ing activity continue to be higher after the halt. During the first 30 minutesafter the halt, market order submissions are nearly 700 percent higher thanon nonhalt days, and both submissions and cancellations of limit orders areover five times nonhalt day levels. Although the abnormal statistics de-crease in magnitude over time, market order submissions, limit order sub-missions, and limit order cancellations all remain unusually high for at leasttwo hours after the halt.

Results for order imbalance halts are provided in Panel B of Table II. Withthe exception of market order cancellations, none of the abnormal measuresis statistically different from zero in periods 24 through 22. Unlike newshalts, however, order imbalance halts are associated with increased tradingactivity in the 30 minutes prior to the halt. Market order submissions in-crease by 569 percent relative to nonhalt days, and limit order submissionsincrease by over 300 percent. In addition, both market order cancellationsand limit order cancellations are substantially higher than on nonhalt days.These results are consistent with the unusual trading conditions that wouldresult in an order imbalance trading halt. Notably, these results suggestthat market conditions change abruptly, providing little time for the special-ist to draw liquidity to the market.

5 The abnormal measure of cancellations is generally equal to 2100 percent in the prehaltand posthalt periods. However, this results from zero cancellations on the halt day and a pos-itive but small mean number of cancellations on nonhalt days. The economic importance ofthese numbers is therefore minimal.


During and after the halt, the results for order imbalance halts are sim-ilar to those for news halts but are much larger in magnitude. Market ordersubmissions are 319 percent higher during the halt, over 1,400 percent higherduring the first posthalt period, and remain significantly higher for up totwo hours after the halt. In addition, market order cancellations during thehalt are over 8,000 percent higher than nonhalt day values. Both limit ordersubmissions and cancellations during the halt are over 1,100 percent higherthan nonhalt day values. As is the case with news halts, limit order activityremains significantly higher than nonhalt day levels for at least two hoursafter the halt.

Together, the results in Table II suggest that investors use the temporarystoppage of trade to cancel old limit orders and reposition their trading in-terests both during and following the halt. The increased activity docu-mented in Table II is also consistent with price discovery. However, we areunable to determine whether price discovery is improved or limited by thetrading halt. Finally, the results suggest that the conditions that precipitatean imbalance halt develop quickly.

B. The Limit Order Book

The results in Section III.A point to a marked increase in order activityduring trading halts and for several periods after trading halts. The in-crease in cancellations during the halt is particularly noticeable. In thissection, we examine whether these changes in order activity have a signif-icant impact on the state of the limit order book around the time of the halt.

We reconstruct the limit order book at 30-minute intervals around thetrading halt using NYSE System Order Data and the methodology of Kava-jecz ~1999!. The basic methodology involves three steps: First, we use execu-tion and cancellation records to identify the set of limit orders outstandingat the beginning of the sample period ~day 210!. This set of orders, referredto as the prebook, includes all limit orders that were canceled or executedduring our sample period and for which the submission date is prior to day210. Second, we update the prebook with all submissions prior to the dateand time of interest. Finally, we remove any orders that were executed orcanceled prior to the date and time of interest. The remaining orders providean approximation of the limit order book at a particular point in time.

At each 30-minute interval, we compute four measures of share depth inthe limit order book: the cumulative depth within $0.25 of the quote mid-point, the cumulative depth within $0.50 of the quote midpoint, the cumu-lative depth within $1.00 of the quote midpoint, and the total depth in thelimit order book. Together, these measures provide a detailed characteriza-tion of the limit order book.6 Results are provided for combined bid-side andask-side depth ~Table III! and separately for bid-side and ask-side depth

6 Goldstein and Kavajecz ~2000a, 2000b! use similar measures to study market depth aroundthe change in tick size on the NYSE and the marketwide circuit breaker on the NYSE, respectively.


Table III

Limit Order Book Depth around Trading HaltsThe sample includes 295 order imbalance and news halts of NYSE-listed common stocks that were called after 10:00 a.m. Limit order book depthis estimated using NYSE System Order Data and the methodology of Kavajecz ~1999!. Statistics are generated for share depth within $0.25 ofthe specialist’s quote midpoint, share depth within $0.50 of the midpoint, share depth within $1.00 of the midpoint, and total share depth. Foreach depth category, the table lists the median depth and the median value of abnormal depth. The abnormal depth measure is defined as thedepth on the halt day minus the mean depth on nonhalt days, stated as a percentage of the nonhalt day mean. Nonhalt days include the 10 daysbefore and 10 days after the halt. For the abnormal measures, ***, **, and * indicate that the median is statistically significant at the 1, 5, and10 percent levels, respectively, based on a signed-rank test. Thirty-minute prehalt periods are measured backward from the beginning of the halt.Thirty-minute posthalt periods are measured forward from the reopening time.

Depth within $0.25 Depth within $0.50 Depth within $1.00 Total Depth

Period N

HaltDay

Depth

AbnormalDepth

~%!

HaltDay

Depth

AbnormalDepth

~%!

HaltDay

Depth

AbnormalDepth

~%!

HaltDay

Depth

AbnormalDepth

~%!

Panel A: News Halts

24 182 11,571 −21.36** 18,370 212.06 29,775 211.21 54,249 −8.40**23 196 9,442 −28.07*** 17,023 −22.31*** 28,900 −17.56*** 51,887 −8.00**22 214 11,350 −27.19*** 20,300 −21.09*** 27,194 −15.77*** 49,400 −8.07***21 250 10,822 −28.08*** 20,000 −18.90*** 29,400 −13.85*** 49,736 −9.70***

Halt 251 10,950 −30.42*** 20,858 −18.76*** 32,360 −12.53*** 55,545 −10.66**Reopen 251 9,206 −42.95*** 29,950 −24.64*** 32,225 −15.20* 62,414 23.31

11 222 11,348 −35.17*** 23,035 −21.40*** 35,185 −14.03** 64,229 3.75*12 184 13,395 −25.60** 28,667 212.95 43,150 20.53 73,157 11.30***13 160 15,012 −14.98* 29,291 22.01 42,543 9.93* 73,849 13.46***14 119 16,743 28.05 31,934 20.02 47,183 3.82 83,212 13.14***


24 25 7,450 216.27 10,400 221.48 13,700 220.37 38,307 219.9723 31 2,500 −60.19*** 10,650 −28.87** 12,593 −30.11** 28,274 −13.72**22 37 5,200 −52.73* 8,300 −48.19* 13,500 226.75 29,400 218.6721 44 2,500 −56.05** 5,700 241.34 10,790 235.54 29,581 218.49

Halt 44 3,700 −55.93* 6,162 −59.31* 11,000 240.33 30,117 6.42Reopen 44 5,650 −73.29** 13,507 243.47 20,849 27.10 81,560 67.85***

11 40 4,900 222.71 16,222 220.55 23,000 27.08** 63,740 78.97***12 37 6,500 247.09 13,925 10.88 25,475 40.04** 89,683 72.76***13 36 6,800 215.93 11,900 14.56 27,798 44.25*** 84,727 62.50***14 31 5,808 222.64 14,672 20.06 23,183 46.53*** 85,561 92.44***

1780T

he

Jou

rnal

ofF

inan

ce

~Table IV!. For the separate analyses of bid and ask-side depth, the sampleis categorized according to the price change during the halt. As in Table II,we calculate abnormal statistics for each time period by comparing the haltday value with the average value across nonhalt days. For each period, thetable lists the median share depth across all halts and the median value ofthe abnormal depth statistic. Statistical tests are again based on a signed-rank test.

Panel A of Table III presents the results for the subsample of news halts.Examining the periods prior to the halt, we see that limit order book depthis substantially lower on the halt day than on nonhalt days. This reductionin depth is most evident near the quotes. At the time of the halt, forexample, depth within $0.25 of the midpoint is 30 percent lower on thehalt day than on nonhalt days, whereas total depth is only 11 percentlower.

At the reopen of trading, depth within $0.25 of the midpoint remains sig-nificantly lower than on nonhalt days. This reduction persists for approxi-mately one hour after the halt. As we move away from the quotes, however,we see a different picture. Share depth within $0.50 of the midpoint andwithin $1.00 of the midpoint is significantly lower than on nonhalt days inperiod 11 but does not differ from nonhalt days thereafter. For total depth,we observe abnormally low depth prior to the halt but abnormally high depthafter the halt. One possible explanation for this difference could be in-creased activity by limit order traders who face substantial uncertainty aboutthe new price. Another possible explanation is that this increased depth re-sults from stale limit orders that do not ref lect new information. We addressthese possibilities below.

Panel B of Table III presents results for order imbalance halts. The resultsare similar to those reported for news halts with the following exceptions.First, depth prior to the halt is only marginally lower than on nonhalt days,with the most significant difference evident 90 minutes prior to the halt.Second, only depth within $0.25 of the midpoint is significantly lower at thereopen. This lack of significance may be the result of the small number oforder imbalance halts in the sample. Despite the sample size, however, totaldepth is significantly higher at the reopen,and both total depth and depthwithin $1.00 of the midpoint are significantly higher than on nonhalt daysin periods 11 through 14.

Separate results for bid-side and ask-side depth are provided in Table IV.Results for positive, negative and zero price change halts are listed in Pan-els A, B, and C, respectively.7 Results for depth near the quotes are similarfor all price change categories and are consistent with the results fromTable III. Both bid-side and ask-side depth near the quotes is significantlylower on the halt day than on nonhalt days, especially in the periods imme-diately before and immediately after the halt. Interestingly, even for zero

7 In Table IV, the samples of order imbalance and news halts are combined due to the smallnumber of order imbalance halts in each price change category.


Table IV

Limit Order Book Depth around Trading Halts (by price change category)The sample includes 295 order imbalance and news halts of NYSE-listed common stocks that were called after 10:00 a.m. Limit order book depthis estimated using NYSE System Order Data and the methodology of Kavajecz ~1999!. Statistics are generated for share depth within $0.25 ofthe specialist’s quote midpoint and share depth within $1.00 of the midpoint. For each depth category, the table lists the median depth and themedian value of abnormal depth and the median value of abnormal depth0volume. The abnormal depth measure is defined as the depth on thehalt day minus the mean depth on nonhalt days, stated as a percentage of the nonhalt day mean. Nonhalt days include the 10 days before and10 days after the halt. For the abnormal measures, ***, **, and * indicate that the median is statistically significant at the 1, 5, and 10 percentlevels, respectively, based on a signed-rank test. Thirty-minute prehalt periods are measured backward from the beginning of the halt. Thirty-minute posthalt periods are measured forward from the reopening time.

Bid-Side Depth Ask-Side Depth

Period N

Halt DayDepth within

$0.25

AbnormalDepth within

$0.25~%!

Halt DayTotal Depth

AbnormalTotal Depth

~%!

Halt DayDepth within

$0.25

AbnormalDepth within

$0.25~%!

Halt DayTotal Depth

AbnormalTotal Depth

~%!

Panel A: Positive Price Change Halts

24 80 5,550 231.71 19,550 −25.87*** 4,750 −43.17*** 27,969 −16.92**23 87 5,060 240.26 18,400 −27.00** 4,700 −48.94*** 28,551 −18.55***22 97 5,700 235.35 18,200 −24.70** 4,650 −40.26*** 28,553 −12.32**21 111 4,562 −46.16** 18,200 −24.94*** 5,750 −44.76*** 29,448 −11.69*

Halt 111 4,800 −42.84* 18,150 217.33 4,213 −47.39*** 29,051 −19.29**Reopen 111 3,475 −54.75*** 41,325 33.19*** 1,850 −85.09*** 24,150 −36.86***

11 98 5,974 −47.16*** 37,887 32.35*** 4,800 −47.52* 32,129 215.5712 82 7,042 234.97 45,275 39.25*** 7,340 218.51 38,197 1.4113 74 6,050 232.24 42,725 36.25*** 10,350 23.33 39,275 4.5914 48 5,700 224.39 42,800 29.78*** 14,000 14.41 47,814 8.11*

1782T

he

Jou

rnal

ofF

inan

ce

Panel B: Negative Price Change Halts

24 92 3,200 −37.57*** 18,250 −19.48** 4,634 −37.18* 28,800 29.6923 101 3,875 −41.03*** 16,890 −21.80*** 2,500 −45.09*** 27,295 27.0822 112 3,100 −46.61*** 17,740 −17.63** 3,000 −42.62** 26,981 211.7521 126 3,600 −46.89*** 21,512 −18.43*** 3,000 −43.59*** 27,000 213.35

Halt 127 3,750 −41.21*** 21,700 216.84 2,450 −58.26*** 28,168 26.29Reopen 127 2,309 −71.21*** 19,002 223.30 1,900 −82.82*** 44,991 15.35***

11 111 3,198 −47.11*** 24,512 226.54 4,200 −51.18*** 39,450 21.39***12 93 3,500 −51.53*** 28,452 215.44 5,000 −41.10*** 47,146 18.60***13 80 3,800 −49.67** 30,300 210.37 5,800 229.85 50,025 20.56***14 67 4,922 234.93 33,599 29.30 5,800 241.01 54,400 29.77***

Panel C: Zero Price Change Halts

24 34 3,100 229.33 18,600 23.39 9,100 53.80** 24,525 8.0023 39 4,320 239.13 17,550 23.12 3,800 22.61 24,512 5.7222 42 5,000 233.03 20,003 23.40 5,100 29.50 23,490 20.4221 57 4,600 −52.14** 17,549 23.78 3,700 219.32 21,350 22.07

Halt 57 3,285 −45.76** 18,600 26.27 4,000 220.32 21,190 22.16Reopen 57 3,000 −61.88*** 25,900 0.34 2,500 −27.92** 25,500 29.45

11 53 5,000 −36.17** 25,800 25.49 3,165 250.44 28,690 25.4412 46 5,150 −31.86** 26,750 1.04 3,450 −64.14** 29,110 27.1113 42 3,500 −39.40* 27,170 3.10 5,250 230.95 34,376 24.3014 35 5,900 231.77 25,100 23.89 6,525 244.53 38,352 27.34

Ord

erF

lowan

dL

iquid

ityarou

nd

NY

SE

Trad

ing

Halts

1783

price change halts, bid-side depth near the quotes appears to be unusuallylow around trading halts. For positive and negative price change halts, thereis also evidence that total depth is low in the prehalt periods relative tononhalt days.

The results for total depth in the posthalt periods provide some insightinto the source of increased depth away from the quotes. For positive pricechange halts, total depth on the bid side is significantly higher than onnonhalt days, whereas total depth on the ask side does not differ from non-halt day values. Just the opposite pattern is evident for negative price changehalts. These results suggest that at least some of the increase in total depthin the posthalt periods may be due to stale limit orders that do not ref lectnew information. As the price moves away from these orders, depth awayfrom the quotes increases but only on one side of the market. The results forzero price change halts are consistent with this explanation, because there isno evidence of an increase in total depth on either the bid or ask side forthese halts.

The results in Table IV suggest that a significant number of stale limitorders may remain on the book after the halt. However, the results inTable II provide evidence of substantial submission and cancellation activ-ity before, during, and after halts. To determine what portion of the limitorder book represents new relative to stale orders, we break down the totallimit-book depth at the reopening time into orders submitted prior to thehalt ~Old! and orders submitted during the halt ~New!. The breakdown ispresented in Figure 1. Panels A, B, and C provide results for zero pricechange halts, positive price change halts, and negative price change halts,respectively.

For positive and negative price change halts, the vast majority of limit-book depth is associated with new orders. In addition, the increased biddepth for positive price change halts and increased ask depth for negativeprice change halts appears to result primarily from new orders submittedduring the halt rather than from old orders that do not ref lect new infor-mation. For positive and negative price change halts, there also appears tobe a significant amount of “new” depth on both sides of the book. Theseresults suggest that, during the halt, the limit order book fills in around anew consensus price.8 The results for zero price change halts provide aninteresting contrast. For these halts, total depth is higher throughout thelimit order book, and the majority of limit-book depth comes from orderssubmitted prior to the halt. These results suggest that information effects,rather than the halt itself, cause traders to cancel old orders and repositiontheir trading interests.

In summary, the results in Tables III and IV suggest that limit orderbook depth is low prior to trading halts, especially near the quotes. After thehalt, depth near the quotes is marginally lower than on nonhalt days, whereasdepth away from the quotes is actually higher than on nonhalt days. Although

8 The relation between the limit order book and reopening price is analyzed in Section IV.


a portion of this increased depth appears to be the result of stale limit or-ders, the results in Figure 1 suggest that the majority of the reopening limitorder book is made up of new orders.

C. Bid-Ask Spreads and Quoted Depth

As an alternative description of liquidity around trading halts, we analyzefour measures of the bid-ask spread: the specialist’s quoted bid-ask spread,the limit order spread, the 5,000-share spread, and the 10,000-share spread.The first measure is the difference between the specialist’s posted ask andbid quotes. The second measure is the difference between the best ask priceand the best bid price in the limit order book. These measures represent thecost of transacting small orders. The final two measures are meant to proxyfor the transactions costs associated with large trades. The 5,000-share spreadis the difference between the 5,000-share ask and the 5,000-share bid, wherethe 5,000-share ask ~bid! is the price one would have to go up to ~down to!to reach 5,000 shares of cumulative depth in the limit order book. The 10,000-share spread is defined similarly for 10,000 shares of cumulative depth.9Using all four measures allows us to examine transaction costs affecting avariety of traders. All spreads are stated as a percentage of the bid-askmidpoint.

The results of the bid-ask spread analysis are provided in Table V. Thetable lists both the median percentage spread on the halt day and the me-dian abnormal spread measure calculated relative to nonhalt days. For newshalts ~Panel A!, the limit order spread in all prehalt periods is significantlylower than on nonhalt days. In addition, both the limit order spread and thespecialist’s quoted spread are unusually low at the time of the halt, thoughthe magnitude of the difference is small for the quoted spread. There is littleevidence of a systematic pattern in either the 5,000-share or 10,000-sharespread prior to the halt. At the reopen, all four spread measures are signif-icantly higher than on nonhalt days. The specialist’s quoted spread is31 percent higher than the nonhalt day value, whereas the differences inlimit-order spreads range from 21 percent higher for the best limit orderspread to eight percent higher for the 10,000-share spread. In the posthaltperiods, there is some evidence that spreads are lower than on nonhalt days,

9 In many cases, the total depth on either the bid or ask side of the limit order book is lessthan 5,000 ~or 10,000! shares. To handle these situations, we define the minimum 5,000-sharebid ~maximum 5,000-share ask! as the quoted bid price minus $2.00 ~quoted ask price plus$2.00!. This results in a maximum 5,000-share spread equal to the quoted spread plus $4.00.Similarly, we define the minimum 10,000-share bid ~maximum 10,000-share ask! as the quotedbid price minus $3.00 ~quoted ask price plus $3.00!. The conclusions are unchanged if minimumbids and maximum asks are defined based on the minimum bid and maximum ask in the limitorder book, respectively. We also reach similar conclusions if the 5,000-share and 10,000-sharespread are defined as missing when the total depth on either side of the book is less than 5,000or 10,000 shares, respectively.


especially for the 5,000-share and 10,000-share spread measures. This re-sult is consistent with the increase in posthalt depth away from the quotes,as discussed in Section III.B.10

10 The distribution of abnormal spreads is skewed. Based on mean abnormal statistics, all spreadmeasures are significantly higher at the halt time, reopen time, and period 11, whereas there islittle evidence that spreads differ relative to nonhalt days for other prehalt and posthalt periods.

Figure 1. Limit order book depth at the reopen. The figures show cumulative limit orderbook depth at various distances away from the bid-ask midpoint. Limit order book depth isestimated using NYSE System Order Data and the methodology of Kavajecz ~1999!. Old rep-resents orders submitted prior to the halt time. New represents orders submitted during thehalt. Panel A provides results for halts that result in no price change. Results for positive pricechange halts and negative price change halts are provided in Panels B and C, respectively.~Figure continues on facing page.!


For imbalance halts ~Panel B!, the results are even more dramatic. Con-sistent with the results for news halts, both the 5,000- and 10,000-sharespreads are significantly lower than nonhalt day values in the posthalt pe-riods. Although there is little evidence of abnormal spreads in periods 24through 21, the quoted spread, the 5,000-share spread and the 10,000-sharespread at the halt time are all significantly wider than on nonhalt days.Most notably, the quoted spread at the time of the halt is 361 percent abovenonhalt levels. In addition, the median quoted spread is 2.33 percent, whereasthe median limit order book spread is only 0.38 percent. These striking re-sults for imbalance halts are consistent with specialists “spreading the quote”to signal the price level necessary to resolve the imbalance. Specifically, ifthe specialist expects that a substantial price move would be necessary toexecute electronic market orders, the specialist may move one side of thequote to indicate the likely execution price. Traders then have the opportu-nity to cancel their market orders in light of this new information.11

To more closely examine the time-series behavior of spreads around trad-ing halts, we calculate average dollar spreads at 15-second intervals prior toand after the halt. Cross-sectional means are calculated separately for newsand for order imbalance halts.12 These average spreads are plotted in Fig-ure 2. For both halt types, dollar spreads average approximately 20 to 25cents during nontrading periods. For news halts, quoted spreads increase

11 During this brief period, the specialist may post quotes that bypass existing limit orders.This is evident in Panel B of Table V, where the quoted spread at the halt time exceeds the bestlimit order spread. Of course, the limit orders would be appropriately filled if an execution wereto occur.

12 Where necessary, intervals from the day prior to the halt and the day after the halt areincluded to ensure that all halts have the same number of prehalt and posthalt observations.

Figure 1. Continued


Table V

Bid-Ask Spreads around Trading HaltsThe sample includes 295 order imbalance and news halts of NYSE-listed common stocks thatwere called after 10:00 a.m. Quoted bid and ask prices are identified using the TAQ database.The limit order book is reconstructed using NYSE System Order Data and the methodology ofKavajecz ~1999!. Quoted spread is the specialist’s quoted bid-ask spread. Limit order spread isthe spread between the best bid and best ask prices in the limit order book. The 5,000-sharespread is the difference between the 5,000-share ask and the 5,000-share bid, where the 5,000-share ask ~bid! is the price associated with 5,000 shares of cumulative depth on the ask ~bid!side of the limit order book. The 10,000-share spread is defined similarly. All spread variablesare measured as a percentage of the bid-ask midpoint. For each spread measure, the table liststhe median spread and the median value of the abnormal spread. The abnormal spread mea-sure is defined as the spread on the halt day minus the mean spread on nonhalt days, statedas a percentage of the nonhalt day mean. Nonhalt days include the 10 days before and 10 daysafter the halt. For the abnormal measures, ***, **, and * indicate that the median is statisti-cally significant at the 1, 5, and 10 percent levels, respectively, based on a signed-rank test.Thirty-minute prehalt periods are measured backward from the beginning of the halt. Thirty-minute posthalt periods are measured forward from the reopening time.

Quoted Spread Limit Order Spread 5,000-Share Spread 10,000-Share Spread

Period

HaltDay

Spread~%!

AbnormalSpread

~%!

HaltDay

Spread~%!

AbnormalSpread

~%!

HaltDay

Spread~%!

AbnormalSpread

~%!

HaltDay

Spread~%!

AbnormalSpread

~%!

Panel A: News Halts

24 0.90 25.23 1.00 −27.84*** 4.54 26.05 9.77 −5.42*23 0.90 28.47 1.00 −28.22*** 4.44 211.82 10.00 0.0522 0.90 27.07 1.13 −16.75** 4.55 −11.30** 10.07 22.5321 0.88 25.65 1.00 −16.30** 4.85 24.78 9.93 23.80

Halt 0.97 −1.02** 1.05 −21.10*** 5.26 0.44 9.73 1.01Reopen 1.03 31.48*** 1.41 21.32*** 5.63 18.55*** 9.35 8.02***

11 0.93 −3.33*** 1.15 211.51 4.96 21.10 8.39 23.3812 0.86 29.36 0.91 −16.54** 3.01 −11.68* 6.54 −9.77**13 0.76 −13.58* 0.88 −17.73*** 3.11 −8.99* 5.53 −20.69***14 0.70 29.42 0.84 −18.16* 2.87 −14.25* 6.46 −14.60**


24 0.62 2.24 0.53 −24.48* 4.39 4.11 9.27 4.7523 0.63 25.02 0.91 3.03 4.86 8.72* 8.77 5.2122 0.61 23.23 0.65 220.82 4.97 1.51 9.48 3.5321 0.52 9.58* 0.81 214.73 5.27 3.64 9.85 3.91

Halt 2.33 361.86*** 0.38 260.33 9.01 33.33*** 12.61 35.13***Reopen 0.76 42.92*** 1.94 54.25*** 4.00 23.39 7.11 25.25

11 0.56 21.57 0.70 225.75 4.03 218.67 6.28 −22.59**12 0.54 20.14 0.81 28.19 3.16 211.48 5.84 −19.94**13 0.50 27.93 0.72 224.43 3.19 −37.98** 4.92 −26.67***14 0.60 9.09 0.61 223.26 3.31 −25.95*** 5.88 −31.24***


slightly prior to the halt, remain higher at the reopen, and quickly return toprehalt levels. However, the most significant results are apparent for orderimbalance halts. Average dollar spreads begin to widen in the minutes priorto imbalance halts, increasing to as much as $1.80. As we discussed above,this result is consistent with specialists “spreading the quote” prior to orderimbalance halts. Interestingly, reopening spreads for order imbalance haltsappear to be very close to normal prehalt levels. For news halts, tests ofmeans and medians suggest that spreads begin to increase approximately 15to 90 seconds prior to the halt and return to prehalt levels ~defined as onehour prior to the halt! within 21 minutes. For order imbalance halts, similartests suggest that spreads begin to increase 12 to 13 minutes prior to thehalt and are back to normal within two minutes of the reopen.

The results in the previous two sections provide evidence that limit order bookdepth is unusually low and bid-ask spreads are unusually high immediatelyafter trading halts. Given this decrease in liquidity, it is interesting to deter-mine whether the specialist and f loor participants step in to provide additionalliquidity around trading halts. To address this question, we examine the quoteddepth ~at the best bid or offer! and the f loor contribution to depth. The f loorcontribution is calculated by subtracting from the quoted depth the total depthavailable from the limit order book at prices at or better than the posted quote.The f loor contribution to depth is presented both in terms of shares contrib-uted and proportion of quoted depth contributed. The results for news and im-balance halts are presented in Panels A and B of Table VI, respectively.

Figure 2. Mean dollar spreads around trading halts. The figure plots the mean dollarspread by halt category at 15-second intervals prior to and after the halt. The halt sampleincludes 295 order imbalance and news halts of NYSE-listed common stocks that were calledafter 10:00 a.m.


For news halts, quoted depth is substantially lower before and for one-halfhour after the trading halt. The quoted depth is 26 percent less than onnonhalt days in period 24, 28 percent less than on nonhalt days at the halt

Table VI

Quoted Depth and Floor Contribution to Depth around Trading HaltsThe sample includes 295 order imbalance and news halts of NYSE-listed common stocks thatwere called after 10:00 a.m. Quoted depth is the sum of the bid and ask depths quoted by thespecialist. The depth contributed by f loor participants is defined as the difference between thequoted depth and the total depth on the limit order book at or inside the quotes. Limit orderbook depth is estimated using NYSE System Order Data and the methodology of Kavajecz~1999!. The percentage of depth contributed by f loor participants is defined as depth contrib-uted by the f loor divided by total quoted depth. The abnormal depth measure is defined as thedepth on the halt day minus the mean depth on nonhalt days, stated as a percentage of thenonhalt day mean. Abnormal measures of f loor contribution are defined similarly. Nonhalt daysinclude the 10 days before and 10 days after the halt. For the abnormal measures, ***, **,and * indicate that the median is statistically significant at the 1, 5, and 10 percent levels,respectively, based on a signed-rank test. Thirty-minute prehalt periods are measured back-ward from the beginning of the halt. Thirty-minute posthalt periods are measured forwardfrom the reopening time.

Quoted DepthFloor Contribution

to Depth% of Depth Contributed

by the Floor

PeriodHalt Day

Depth

AbnormalDepth

~%!Halt Day

Depth

AbnormalDepth

~%!Halt Day

Depth

AbnormalDepth

~%!

Panel A: News Halts

24 7,550 −26.14** 1,422 −56.78*** 33.02 215.3523 7,400 −39.10*** 1,256 −61.46*** 31.43 217.9222 8,000 −30.74*** 1,250 −52.85*** 33.33 211.0421 8,000 −28.06*** 1,200 −52.40*** 32.33 29.66

Halt 6,900 −28.68*** 1,000 −55.75*** 28.57 223.81Reopen 9,000 −25.45*** 3,000 233.07 50.82 36.12***

11 9,000 −35.13*** 2,000 236.13 48.68 21.38***12 11,500 215.33 2,915 227.17 40.01 14.37*13 11,000 20.98 2,100 235.90 34.64 3.0214 10,000 212.21 2,000 247.72 28.57 220.32


24 4,000 −43.83*** 1,750 −61.83*** 36.55 219.0623 6,000 −43.12** 2,000 −34.39** 50.00 14.1922 5,500 −23.08** 1,500 −62.72* 41.53 26.4021 4,800 −37.68* 1,500 −56.99* 61.82 8.26

Halt 10,100 24.61 2,350 251.50 72.35 26.15Reopen 6,500 224.31 3,350 214.04 70.38 21.61**

11 6,000 216.07 2,228 227.64 50.00 26.7812 9,700 213.87 4,000 29.25 56.43 30.5013 7,000 224.17 2,250 27.79 50.00 0.0914 8,750 20.21 3,200 224.28 50.00 4.80


time, and 25 percent less than on nonhalt days at the reopen. A similarpattern is observed for the f loor contribution to depth, except at the reopenand just after the halt, where the f loor contribution to depth does not differsignificantly from nonhalt day values. These results are generally consis-tent with the decreased limit order book depth near the quote documented inSection III.B. However, results based on the proportion of depth contributedby the f loor lead to a substantially different conclusion. At the reopen andfor the next hour, the proportion of depth provided by the f loor actuallyincreases. For example, the proportion of depth contributed by f loor partici-pants is 36 percent higher than nonhalt day values at the reopen and re-mains 21 percent higher than normal after 30 minutes. The results for orderimbalance halts ~Panel B! are generally similar, though the significant in-crease in proportion of depth contributed is evident only at the reopen. To-gether, the results in Table VI suggest that f loor participants, including thespecialist, step in to provide liquidity following trading halts.

IV. The Limit Order Book and Reopening Prices

The results in Section III suggest that at least some traders repositiontheir trading interest around a new consensus price. As an additional test,we examine the informativeness of prices in the limit order book. Specifi-cally, we follow Madhavan and Panchapagesan ~2000! in calculating the pricesthat most closely equate supply and demand across all limit orders ~thelimit-book clearing price! and all market and limit orders ~the system-clearing price!. We then evaluate the informativeness of these prices by mea-suring the extent to which these prices can predict subsequent prices.13

This analysis is useful for a number of reasons. First, system-clearing andlimit-book clearing prices may not exist if there is a substantial order im-balance. Thus, the existence of market-clearing prices provides one indica-tion that a consensus price has been reached. Second, if the specialist hasbeen successful at attracting depth around a new price, then the prices thatclear supply and demand should be close to the new security value. On theother hand, if there is a wide gap between supply and demand, clearingprices will not be good estimates of future prices.

We measure the informativeness of market-clearing prices by testing theproportion of the total price change that is explained by the clearing price.In other words, a good price should explain a large proportion of the pricechange resulting from the halt. In addition, we follow Madhavan and Pan-chapagesan ~2000! in testing whether the specialist’s reopening price pro-vides information beyond that contained in the clearing prices. This analysis

13 This analysis focuses on price discovery during unscheduled market closures ~trading halts!.Several recent papers examine price discovery during scheduled market closures. These includeMadhavan and Panchapagesan ~2000! for the morning open on the NYSE, Biais, Hillion, andSpatt ~1999! for the morning open on the Paris Bourse, and Cao, Ghysels, and Hatheway ~2000!for the morning open on Nasdaq.


provides information about whether the specialist reacts to liquidity in thelimit order book or uses additional information beyond the limit order bookwhen setting the reopening price. We employ a two-stage regression ap-proach, specified as follows:

Stage 1: ln~Pi, post 0Pi, pre! 5 a 1 b ln~Pi,1 0Pi, pre! 1 ei , ~2!

Stage 2: [ei 5 g 1 d ln~Pi,2 0Pi, pre! 1 ui . ~3!

We define Pi, pre as the price one hour before halt i and Pi, post as the priceone hour after halt i. These prices proxy for the initial equilibrium price andthe new equilibrium price, respectively. Prices Pi,1 and Pi,2 are equal to thesystem-clearing price, the limit-book clearing price, or the specialist’s re-opening price for halt i, where the limit-book ~system! clearing price is theprice that most closely equates supply and demand across all limit orders~market and limit orders! just prior to the reopen. Thus, in the first-stageregression we evaluate the quality of a given price ~Pi,1! and in the secondstage we see what incremental value is contained in an alternative price~Pi,2!. If Pi,1 is an unbiased predictor of the new equilibrium price and Pi,2provides no additional information, then b will equal one and d will equalzero. If, on the other hand, Pi,2 ref lects some information not captured inPi,1, then d will be statistically different from zero.

The regression results are provided in Table VII. We find that a limit-bookclearing price exists for every halt and a system-clearing price exists for allbut two halts. Thus, extreme imbalances appear to have been all but elim-inated. Most important, the clearing prices are good estimates of the sub-sequent equilibrium price. On average, the system-clearing price ref lects 81percent of the total price change, and the limit-book clearing price ref lects82 percent of the total price change. As in Madhavan and Panchapagesan~2000!, the specialist’s reopening price adds incremental explanatory power.For example, the specialist’s reopening price has explanatory power for theresiduals based on the system-clearing price, but the system-clearing priceprovides no additional information beyond that contained in the specialist’sreopening price.

Together, the results in Table VII suggest that significant imbalances areeliminated prior to the reopening trade and that the clearing price is a goodpredictor of the future price. In addition, the specialist appears to use bothclearing prices and supplemental information, such as interest from the trad-ing f loor, when setting the reopening price.

V. The Limit Order Book and Volatility

A. Abnormal Volatility

Lee et al. ~1994! find that several measures of volatility are unusuallyhigh after NYSE trading halts. The temporary decrease in posthalt liquiditydocumented in Section III provides a possible explanation for this increase


in volatility. To formally address this possibility, we calculate abnormal mea-sures of volatility similar to those analyzed in Lee et al. ~1994!. We then testwhether these measures are related to the changes in liquidity documentedin Section III.

For each time period, we calculate abnormal statistics for three measuresof volatility: the number of quote revisions, the absolute value of the trans-action price return, and the high–low transaction price range. For complete-ness, we also calculate the abnormal trading volume. Pre- and posthalt periodsare defined as in the previous section.

Summary statistics for the abnormal volume and volatility measures aredescribed in Table VIII. For each time period, the table lists the medianvalue of the variable and the median value of the abnormal statistic acrossall halts in the sample. The results are consistent with the increase in vol-

Table VII

Informativeness of Opening PricesThe table lists results from the following two-stage regression analysis:

Stage 1: ln~Pi, post 0Pi, pre! 5 a 1 b ln~Pi,1 0Pi, pre! 1 ei ,

Stage 2: [ei 5 g 1 dln~Pi,2 0Pi, pre! 1 ui ,

where Pi, post is the price one hour after halt i, Pi, pre is the price one hour before halt i, Pi,1 andPi,2 are equal to the system-clearing price, the limit-book clearing price, or the specialist’sreopening price, and êi are the residuals from the first regression. The sample includes 295order imbalance and news halts of NYSE-listed common stocks that were called after 10:00 a.m.The limit order book is reconstructed using NYSE System Order Data and the methodology ofKavajecz ~1999!. The limit-book ~system! clearing price is defined as the price that most closelyequates supply and demand across all limit orders ~market and limit orders!. The numberslisted in parentheses below the parameter estimates are p-values. Limit-book clearing pricesexist for all of the sample halts, and system-clearing prices exist for 293 of the sample halts.

First-Stage Regression Second-Stage Regression

[a Zb Adj. R2 ~%! [g Zd Adj. R2 ~%!

Pi,1 5 system-clearing price, Pi,2 5 opening price20.002 0.81 57.2 0.001 0.35 22.5~0.55! ~0.00! ~0.82! ~0.00!

Pi,1 5 opening price, Pi,2 5 system-clearing price20.002 1.06 86.3 20.001 20.001 0.3~0.21! ~0.00! ~0.99! ~0.95!

Pi,1 5 limit-book clearing price, Pi,2 5 opening price20.002 0.82 57.6 0.006 0.35 21.9~0.56! ~0.00! ~0.82! ~0.00!

Pi,1 5 opening price, Pi,2 5 limit-book clearing price20.002 1.06 86.3 20.001 20.003 20.3~0.21! ~0.00! ~0.99! ~0.90!


ume and volatility documented by Lee et al. ~1994!. We find that tradingvolume in the prehalt period is comparable to that on nonhalt days. How-ever, volume is nearly three times greater than nonhalt levels in the post-halt period. In fact, during the 30 minutes after the halt, volume is over 600percent higher than on nonhalt days. In general, most of the volume in-crease occurs in the first 90 minutes after the halt, but trading volumeremains abnormally high for at least two hours after the halt.

The results in Table VIII also confirm that volatility after trading halts issubstantially higher than during comparable periods on nonhalt days. Dur-ing the first 30 minutes after the halt, volatility ranges from three to sixtimes that on nonhalt days. Consistent with the results of Lee et al. ~1994!,abnormal volatility is most significant immediately after the halt and di-minishes over the next two hours. We also find that volatility is higher inthe prehalt period, though the magnitude of the difference between halt andnonhalt days is not as great as for the posthalt period.

Table VIII

Abnormal Volume and Volatility MeasuresThe sample includes 295 order imbalance and news halts of NYSE-listed common stocks thatwere called after 10:00 a.m. Volume and volatility measures are estimated using trade andquote data from the TAQ database. Volume is the total share volume during the period. Abso-lute Return is the absolute value of the return from the last transaction price of the previousperiod to the last transaction price of the current period. Price Range is the difference betweenthe highest and lowest transaction prices during the period. Quote Revisions is defined as thenumber of quote changes during the period for which either the bid or ask prices changed. Foreach variable, the table lists the median value of the variable and the median value of theabnormal measure. The abnormal volume measure is defined as the volume on the halt dayminus the mean volume on nonhalt days, stated as a percentage of the nonhalt day mean.Abnormal return, range, and quote revisions are defined similarly. Nonhalt days include the10 days before and 10 days after the halt. For the abnormal measures, ***, **, and * indicatethat the median is statistically significant at the 1, 5, and 10 percent levels, respectively, basedon a signed-rank test. Thirty-minute prehalt periods are measured backward from the halttime to the opening time. Thirty-minute posthalt periods are defined forward from the reopentime to the close of trading.

VolumeQuote

RevisionsAbsoluteReturn

TransactionPrice Range

Period MedianAbnormal

~%! MedianAbnormal

~%! MedianAbnormal

~%! MedianAbnormal

~%!

Full periodsPre-halt 39,700 212.94 15 10.26*** 1.19 42.86*** 0.375 22.33***Post-halt 140,150 279.10*** 25 272.09*** 1.44 114.99*** 0.625 167.42***

30-minute post-halt periods11 49,800 618.82*** 12 535.79*** 1.09 285.38*** 0.375 500.00***12 26,150 213.26*** 6 192.68*** 0.56 96.78*** 0.125 150.00***13 18,500 121.65*** 5 124.01*** 0.39 38.27*** 0.125 86.15***14 14,500 77.24*** 4 61.29*** 0.26 −5.55* 0.125 33.33***


B. Volatility and Liquidity

As a first look at the relation between volatility and liquidity followinghalts, we provide a graph of abnormal volatility and abnormal limit-bookdepth, where depth is measured at the beginning of each period. For eachposthalt period, Figure 3 shows three measures of abnormal volatility ~thebars! and the abnormal depth within $0.25 of the quote midpoint ~the line!.To limit scale differences, the abnormal volatility measures correspond tothe left-hand Y-axis and the abnormal depth measure corresponds to theright-hand Y-axis. From Figure 3, it appears that there is a negative rela-tion between abnormal volatility and abnormal depth.

To test the relation between abnormal volatility and abnormal depth, wefirst compare median abnormal volatility across groups ranked by abnormaldepth. For each posthalt period, we divide the sample into two equal groups:those with abnormal depth above the median and those with abnormal depth

Figure 3. Abnormal volatility and limit order book depth. The figure plots median valuesof abnormal depth and abnormal volatility for 30-minute periods after trading halts. Abnormalvolatility ~the bars! corresponds to the left-hand Y-axis. Abnormal depth ~the line! correspondsto the right-hand Y-axis. The halt sample includes 295 order imbalance and news halts ofNYSE-listed common stocks that were called after 10:00 a.m. Absolute return is the absolutevalue of the return from the last transaction price of the previous period to the last transactionprice of the current period. High–low price range is the range between the highest and lowesttransaction prices during the period. Number of quote revisions is defined as the number ofquote changes during the period for which either the bid or ask price changed. Limit order bookdepth is estimated using NYSE System Order Data and the methodology of Kavajecz ~1999!.Depth is measured within $0.25 of the specialist’s quote midpoint. Abnormal depth is defined asthe percentage differences between the halt day value and the nonhalt day mean. Abnormalvolatility measures are defined similarly.


below the median. We then test whether median abnormal volatility differsacross depth categories. Rankings are based on two alternative measures ofabnormal depth, and tests are performed for all three measures of abnormalvolatility. Statistical differences are based on the Wilcoxon rank-sum test.

The results of this analysis are provided in Table IX. The first result thatis evident from the table is that the ranking produces a wide variation inabnormal depth. In all posthalt periods, abnormal depth differs significantlybetween the low-depth and high-depth groups at the one percent level. Forperiod 11, there also appears to be a significant relation between abnormaldepth and abnormal volatility. All measures of abnormal volatility duringthis period are higher in the low-depth group than in the high-depth group,and the difference is significant at the five percent level in all but one case.For periods 12 through 14, however, we find little evidence of a significantrelation between abnormal volatility and abnormal depth.

The results in Table IX suggest that the high levels of volatility observedimmediately after trading halts may result, at least partially, from changesin liquidity around the halt. However, these tests do not control for infor-mation effects and the unusually high volume on trading-halt days. To con-trol for these effects, we estimate ordinary least squares ~OLS! regressionsof abnormal volatility on abnormal depth and several control variables. Basedon the results in Table IX, we estimate regressions for period 11 only. Theregressions take the following general form:

Abnormal Volatilityi 5 a0 1 a1 * AbnDepthi 1 a2 * HaltReti

1 a3 * AbnVolumei 1 ei . ~4!

The dependent variables are the abnormal measures of quote revisions,absolute transaction price return, and transaction price range defined above.The explanatory variables include the absolute return during the halt~HaltRet!, the abnormal volume during period 11 ~AbnVolume!, and twoalternative measures of abnormal depth: depth within $0.25 of the quotemidpoint and depth within $0.50 of the midpoint. If the increase in abnor-mal volatility is related to the decrease in liquidity, we would expect thecoefficient on abnormal depth to be negative. If large price changes result inincreased price uncertainty, we would expect HaltRet to be positively asso-ciated with abnormal volatility. Finally, to the extent that abnormal volumeref lects differences of opinion among investors, we expect abnormal volumeto be positively associated with abnormal volatility. The intercept from theregression can be interpreted as the average level of abnormal volatility thatcannot be explained by the control variables.

The regression results are provided in Table X. The adjusted R2 is re-ported next to each equation, and p-values are reported in parentheses un-der each coefficient. In addition, the mean value of each abnormal volatility


Table IX

Abnormal Volatility Measures by Depth CategoryThe sample includes 295 order imbalance and news halts of NYSE-listed common stocks that were called after 10:00 a.m. In addition, the testsinclude one observation for each 30-minute posthalt period up to two hours after the halt. Volatility measures are estimated using trade andquote data from the TAQ database. Absolute Return is the absolute value of the return from the last transaction price of the previous period tothe last transaction price of the current period. Price Range is the difference between the highest and lowest transaction prices during the period.Quote Revisions is defined as the number of quote changes during the period for which either the bid or ask prices changed. Limit order bookdepth is estimated using NYSE System Order Data and the methodology of Kavajecz ~1999!. Depth is measured within $0.25 of the specialist’squote midpoint and within $0.50 of the quote midpoint. Abnormal depth is defined as the depth on the halt day minus the mean depth on nonhaltdays, stated as a percentage of the nonhalt day mean. Abnormal measures of volatility are defined similarly. To perform the tests, observationsare ranked according to the depth measure and grouped into two categories based on this ranking. Medians for the high and low depth categoriesare then compared using a Wilcoxon Rank Sum test. ***, **, and * indicate that medians are statistically different at the 1, 5, and 10 percentlevels, respectively.

Categorized by Depthwithin $0.25 of the Midpoint

Categorized by Depthwithin $0.50 of the Midpoint

Period

AbnormalDepth

~%!

AbnormalQuote

Revisions~%!

AbnormalReturn

~%!

AbnormalPriceRange

~%!

AbnormalDepth

~%!

AbnormalQuote

Revisions~%!

AbnormalReturn

~%!

AbnormalPriceRange

~%!

Period 11Low depth 285.82 652.69 304.08 620.00 268.75 666.37 322.38 589.66High depth 21.42*** 459.83** 229.30 370.59** 20.45*** 455.56*** 203.80** 370.59**

Period 12Low depth 271.17 224.56 114.04 150.00 250.27 185.71 90.52 150.00High depth 24.54*** 182.86* 83.92 116.67 32.28*** 201.08 107.91 150.00

Period 13Low depth 267.39 102.19 48.27 100.00 259.01 119.18 55.29 100.00High depth 38.74*** 148.08 17.45 81.82 36.52*** 138.10 26.04 58.00

Period 14Low depth 259.06 57.57 3.79 41.60 241.42 53.85 20.35 27.66High depth 29.88*** 93.55 23.06 27.02 37.50*** 98.35 0.56 42.86

Ord

erF

lowan

dL

iquid

ityarou

nd

NY

SE

Trad

ing

Halts

1797

Table X

Abnormal Volatility Regressions for Period +1The table lists coefficients from a regression of period 11 abnormal volatility on several ex-planatory variables. The sample includes 295 order imbalance and news halts of NYSE-listedcommon stocks that were called after 10:00 a.m. Period 11 extends from the reopen until 30minutes after the reopen. The regression takes the following general form:

Abnormal Volatilityi 5 a0 1 a1 * AbnDepthi 1 a2 * HaltReti 1 a3 * AbnVolumei 1 ei

The dependent variables are abnormal measures of quote revisions, absolute transaction pricereturn, and transaction price range. The number of quote revisions is defined as the number ofquote changes during the period for which either the bid or ask price changed. Absolute returnis the absolute value of the return from the last transaction price of the previous period to thelast transaction price of the current period. High–low price range is the range between thehighest and lowest transaction prices during the period. The explanatory variables are abnor-mal depth, absolute halt return, and abnormal volume. Limit order book depth is estimatedusing the methodology of Kavajecz ~1999!. The absolute return during the halt is defined as theabsolute percentage return from the last price prior to the halt to the reopening price. Volumeis the total share volume during the period. All abnormal measures are defined as the percent-age difference between the halt day value and the nonhalt day mean, where nonhalt daysinclude the 10 days before and 10 days after the halt. Panel A lists results based on depthwithin $0.25 of the quote midpoint. Panel B lists results based on depth within $0.50 of thequote midpoint. The numbers listed in parentheses below the coefficients are p-values. Themean value of each abnormal volatility measure is provided for comparison. Statistical tests arebased on White’s corrected x2-test if homoskedasticity is rejected at the 5 percent level.

AbnormalVolatilityMeasure

MeanAbnormalVolatility

Intercept~a0!

AbnormalDepth~a1!

AbsoluteHalt Return

~a2!

AbnormalVolume

~a3! Adj. R2

Panel A: Depth within $0.25 of the quote midpoint

Abnormal quote 732.263 699.076 20.591 — — 0.0016revisions ~0.000! ~0.235!

443.157 20.623 36.940 0.089 0.3538~0.000! ~0.122! ~0.000! ~0.000!

Abnormal absolute 468.946 468.481 0.156 — — 20.0034return ~0.000! ~0.701!

432.816 0.198 9.676 0.002 20.0053~0.000! ~0.628! ~0.241! ~0.810!

Abnormal price 867.732 831.356 21.608 — — 0.0086range ~0.000! ~0.082!

653.859 21.369 46.099 0.014 0.0362~0.000! ~0.135! ~0.006! ~0.440!

Panel B: Depth within $0.50 of the quote midpoint

Abnormal quote 732.263 704.536 20.875 — — 0.0069revisions ~0.000! ~0.098!

445.257 20.982 38.026 0.088 0.3613~0.000! ~0.027! ~0.004! ~0.000!

Abnormal absolute 468.946 467.248 0.280 — — 20.0023return ~0.000! ~0.517!

432.017 0.282 9.322 0.002 20.0046~0.000! ~0.515! ~0.256! ~0.801!

Abnormal price 867.732 853.976 21.619 — — 0.0093range ~0.000! ~0.073!

664.559 21.607 48.231 0.013 0.0404~0.000! ~0.071! ~0.004! ~0.437!


measure is provided for comparison. Panel A lists results based on depthwithin $0.25 of the quote midpoint, and Panel B lists results based on depthwithin $0.50 of the midpoint.

The results provide weak evidence of a relationship between abnormaldepth and abnormal volatility. However, the control variables explain onlya small portion of the observed abnormal volatility after halts. As ex-pected, the coefficients on absolute halt return and abnormal volume aregenerally positive, though neither variable is significant in the absolute-return regression and only the halt return is significant in the price-rangeregression.

Examining depth within $0.25 of the quote midpoint ~Panel A!, we findthat abnormal depth is negatively related to abnormal quote revisions andabnormal price range. However, the coefficient on abnormal depth is signif-icant at the five percent level in only one of the six specifications. The re-sults based on depth within $0.50 of the midpoint ~Panel B! are somewhatstronger. The coefficient on abnormal depth within $0.50 is negative andmarginally significant in four of the six specifications. Most important, af-ter controlling for the halt return, abnormal volume, and abnormal depth,the intercept remains significant in every specification. In addition, the mag-nitude of the intercept is similar to the magnitude of the mean abnormalvolatility, suggesting that the model explains, at best, a small portion of theobserved abnormal volatility.14

Taken together, the results in this section provide little support for thehypothesis that abnormal posthalt volatility is explained by decreased li-quidity after the halt. It appears that other factors, such as unmeasuredinformation effects or the closure of trading, explain the majority of theobserved posthalt volatility.

VI. Conclusion

On the NYSE, trading halts can be called for individual securities whensignificant news is expected to impact prices or when large imbalances existbetween buy and sell orders. These halts allow investors to react to newinformation and allow the specialist to determine the new equilibrium pricein the absence of trading. However, some argue that trading halts inhibit thenormal price discovery process and may result in less efficient prices. Weaddress these issues by providing a detailed analysis of order f low and li-quidity around NYSE trading halts. Our results provide important informa-tion related to the reactions of traders to halts and the resulting effects onliquidity, prices, and transaction costs.

14 Conclusions regarding the explanatory power of abnormal depth are similar if either ab-normal volume or the halt return is dropped from the regression. In addition, we reestimatedthe regressions including the duration of the halt as an additional explanatory variable. Thelength of the halt does not appear to affect abnormal volatility after the halt.


We find that submissions and cancellations are extremely high during andimmediately after trading halts. We also find that the limit order book atthe reopen consists primarily of orders submitted during the halt as opposedto stale orders. Consistent with the NYSE’s objectives, these results suggestthat investors use the halt as a chance to cancel old orders and repositiontheir trading interests around the new price. However, we also find evidencethat traders are reluctant to provide liquidity during these unusual marketconditions; limit order book depth near the quotes is unusually low before,during, and immediately after trading halts. During these periods, the spe-cialist and0or floor traders appear to provide additional liquidity to the market.

For news halts, spreads are unusually wide at the reopen and return tonormal levels within 20 to 30 minutes of the halt. These results are consis-tent with specialists setting wide quotes during unusual market conditions.In contrast, bid-ask spreads tend to increase dramatically in the 15 minutesprior to order imbalance halts, returning to normal levels almost immedi-ately after the halt. These results are consistent with specialists “spreadingthe quote” prior to order imbalance halts in order to convey informationabout the imbalance to market participants.

Consistent with Lee et al. ~1994!, we find significant increases in volatil-ity subsequent to trading halts. We find some evidence that this increasedvolatility is related to decreased liquidity around the trading halt. However,decreased limit order book depth appears to explain, at best, a small portionof the increased volatility. The remaining volatility may be the result ofunmeasured information effects or the market closure itself.

In contrast to the limit order activity we observe for individual-securityhalts, Goldstein and Kavajecz ~2000b! find that limit order traders eitherremained inactive or withdrew liquidity during the marketwide circuit breakeron October 27, 1997. In addition, they find that liquidity was substantiallyreduced on the day following the circuit breaker, whereas we find that li-quidity returns to normal rather quickly after individual-security halts. Thesedifferences may be related to the degree of price uncertainty associated withthese events or to institutional features that differ in the two situations.Clearly, additional study comparing these types of events is needed.

REFERENCES

Amihud, Yakov, and Haim Mendelson, 1987, Trading mechanisms and stock returns: An em-pirical investigation, Journal of Finance 42, 533–553.

Biais, Bruno, Pierre Hillion, and Chester Spatt, 1995, An empirical analysis of the limit orderbook and the order f low in the Paris Bourse, Journal of Finance 50, 1655–1689.

Biais, Bruno, Pierre Hillion, and Chester Spatt, 1999, Price discovery and learning during thepreopening period in the Paris Bourse, Journal of Political Economy 107, 1218–1248.

Brown, David P., and Robert H. Jennings, 1989, On technical analysis, Review of FinancialStudies 2, 527–552.

Cao, Charles, Eric Ghysels, and Frank Hatheway, 2000, Price discovery without trading: Evi-dence from the Nasdaq pre-opening, Journal of Finance, forthcoming.

Christie, William G., Shane A. Corwin, and Jeffrey H. Harris, 2000, Et tu, Brute? The role andimpact of trading halts in the Nasdaq stock market, Working paper, Vanderbilt University.


Corwin, Shane A., 1999, Differences in trading behavior across NYSE specialist firms, Journalof Finance 54, 721–745.

Gerety, Mason S., and J. Harold Mulherin, 1994, Price formation on stock exchanges: Theevolution of trading within the day, Review of Financial Studies 7, 609–629.

Goldstein, Michael A., and Kenneth A. Kavajecz, 2000a, Eighths, sixteenths and market depth:Changes in tick size and liquidity provision on the NYSE, Journal of Financial Economics,forthcoming.

Goldstein, Michael A., and Kenneth A. Kavajecz, 2000b, Liquidity provision during circuit break-ers and extreme market movements, Working paper, NYSE.

Greenwald, Bruce C., and Jeremy C. Stein, 1988, The task force report: The reasoning behindthe recommendations, Journal of Economic Perspectives 2, 3–23.

Greenwald, Bruce C., and Jeremy C. Stein, 1991, Transactional risk, market crashes, and therole of circuit breakers, Journal of Business 64, 443–462.

Grossman, Sanford J., 1990. Introduction to NBER symposium on the October 1987 crash,Review of Financial Studies 3, 1–3.

Grundy, Bruce D., and Maureen McNichols, 1989, Trade and revelation of information throughprices and direct disclosure, Review of Financial Studies 2, 485–526.

Kavajecz, Kenneth A., 1999, The specialist’s quoted depth and the limit order book, Journal ofFinance 54, 747–771.

Kodres, Laura E., and Daniel P. O’Brien, 1994, The existence of Pareto superior price limits,American Economic Review 84, 919–932.

Lee, Charles M. C., Mark J. Ready, and Paul J. Seguin, 1994, Volume, volatility, and New YorkStock Exchange trading halts, Journal of Finance 49, 183–214.

Madhavan, Ananth, and Venkatesh Panchapagesan, 2000, Price discovery in auction markets:A look inside the black box, Review of Financial Studies, forthcoming.

New York Stock Exchange, Inc., 1999, Maintaining a fair and orderly market: The critical roleof trading halts and delays, Exchange, July, 1–3.

Stoll, Hans R., and Robert E. Whaley, 1990, Stock market structure and volatility, Review ofFinancial Studies 3, 37–71.

Discussion

DANIEL G. WEAVER*

I enjoyed reading this paper. It fills an important gap in the market micro-structure literature. In recent years, the NYSE and Nasdaq have debatedthe circumstances under which trading in a security would be halted. On theNYSE, the debate has centered on marketwide trading halts and how toreopen trading after a halt. On Nasdaq, discussions have occurred as towhether trading should halt if markets are crossed ~i.e., the highest bid isgreater than the lowest ask! or locked ~i.e., the bid is equal to the ask!.Understanding the mechanics of trading halts, including order submissionstrategies of market participants, and identifying which types of traders

* Zicklin School of Business, Baruch College.


order flow and liquidity around nyse trading halts · 3 news-related trading halts are also...

Documents